惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

GbyAI
GbyAI
爱范儿
爱范儿
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
月光博客
月光博客
腾讯CDC
Last Week in AI
Last Week in AI
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
博客园_首页
量子位
博客园 - 聂微东
Jina AI
Jina AI
小众软件
小众软件
The Cloudflare Blog
有赞技术团队
有赞技术团队
V
V2EX
博客园 - 司徒正美
Apple Machine Learning Research
Apple Machine Learning Research
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
大猫的无限游戏
大猫的无限游戏
博客园 - 三生石上(FineUI控件)
WordPress大学
WordPress大学
阮一峰的网络日志
阮一峰的网络日志
B
Blog
MongoDB | Blog
MongoDB | Blog
L
LangChain Blog
宝玉的分享
宝玉的分享
C
Check Point Blog
H
Hackread – Cybersecurity News, Data Breaches, AI and More
IT之家
IT之家
N
Netflix TechBlog - Medium
I
InfoQ
J
Java Code Geeks
S
SegmentFault 最新的问题
V
Visual Studio Blog
Microsoft Security Blog
Microsoft Security Blog
博客园 - 叶小钗
D
DataBreaches.Net
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
B
Blog RSS Feed
S
Schneier on Security
Webroot Blog
Webroot Blog
P
Proofpoint News Feed
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
T
Threatpost
Project Zero
Project Zero
Scott Helme
Scott Helme
C
CERT Recently Published Vulnerability Notes
P
Privacy International News Feed
T
The Exploit Database - CXSecurity.com
D
Darknet – Hacking Tools, Hacker News & Cyber Security

Shiroha白羽的博客

Golang 踩坑 —— interface 为参数的时候传 nil 指针 Codeforces Round 925 (Div. 3) Codeforces Round 924 (Div. 2) Codeforces Round 923 (Div. 3) Codeforces Round 922 (Div. 2) Codeforces Round 921 (Div. 2) Educational Codeforces Round 161 (Rated for Div. 2) Codeforces Round 920 (Div. 3) Codeforces Round 919 (Div. 2) Hello 2024 Good Bye 2023 Codeforces Round 918 (Div. 4) 个人备份的常用 macOS 清理命令 Codeforces Round 917 (Div. 2) Pinely Round 3 (Div. 1 + Div. 2) Educational Codeforces Round 160 (Rated for Div. 2) Codeforces Round 915 (Div. 2) Codeforces Round 914 (Div. 2) Codeforces Round 913 (Div. 3) Educational Codeforces Round 159 (Rated for Div. 2) Codeforces Round 912 (Div. 2) Codeforces Round 911 (Div. 2) CodeTON Round 7 (Div. 1 + Div. 2, Rated, Prizes!) Educational Codeforces Round 158 (Rated for Div. 2) Codeforces Round 910 (Div. 2) Codeforces Round 909 (Div. 3) Codeforces Round 908 (Div. 2) Educational Codeforces Round 157 (Rated for Div. 2) C++自定义的字面量 Codeforces Round 907 (Div. 2) Codeforces Round 916 (Div. 3) 关于 LRU map 的一些灵感 2023 杭州站 ICPC 现场赛 反复横跳的 Clang-Tidy(cert-dcl21-cpp) Codeforces Round 906 (Div. 2) 一段奇怪的 CPP 代码 Codeforces Round 905 (Div. 3) Codeforces Round 904 (Div. 2) Codeforces Round 903 (Div. 3) Educational Codeforces Round 156 (Rated for Div. 2) Codeforces Round 902 (Div. 2, based on COMPFEST 15 - Final Round) Codeforces Round 901 (Div. 2) Codeforces Round 900 (Div. 3) Codeforces Round 899 (Div. 2) Educational Codeforces Round#155 (Div. 2) Codeforces Round 898 (Div. 4) CodeTON Round 6 (Div. 2) Codeforces Round 897 (Div. 2) Codeforces Round 896 (Div. 2) Codeforces Round 887 (Div. 2) Codeforces Round 895 (Div. 3) 左值-右值-将亡值 blog.mauve.icu Pinely Round 2 (Div. 1 + Div. 2) Harbour.Space Scholarship Contest 2023-2024 (Div. 1 + Div. 2) Codeforces Round 894 (Div. 3) Codeforces Round 888 (Div. 3) Educational Codeforces Round#153 (Div. 2) Codeforces Round 893 (Div. 2) OTPAUTH,两步验证中的通用协议 Codeforces Round 892 (Div. 2) Codeforces Round 891 (Div. 3) Codeforces Round 890 (Div. 2) Educational Codeforces Round#152 (Div. 2) blog.mauve.icu Java Script 的 null 和 undefined 随想 记一次 SQL LEFT JOIN 没有得到预期结果的错误 Codeforces Round#789(Div. 2) GCC/G++ 预编译头性能优化 使用 Junit5 和 Mockito 实现 SpringBoot 的单元测试最优美的解决方案 centOS 防火墙 docker-compse 的问题 C++ 语言实现动态变化的线程池 Codeforces Round#744 (Div. 3) 计算机图形学 Windows 通过网络访问 WSL2 原生 JavaScript 实现图片裁剪 面试复习(计算机图形学) 面试复习(算法) Codeforces Round#706(Div. 2)-Let's Go Hiking 面试复习(Java) 面试复习(Git) 面试复习(Linux) 面试复习(数据库) 面试复习(计算机网络) 面试复习(操作系统) 面试复习(C++) Codeforces Round#699 (Div. 2) 清理 WSL2 的磁盘占用 Codeforces Round#697 (Div. 3) Windows 下的 NTFS 驱动器索引 BUG 计算机网络复习 记一次 Navicat 连接 MySQL 一直报认证错误(Access denied) 计算机网络实验复习 WSL1 使用 Docker 一直无法启动 2020牛客暑期多校训练营(第三场)D-Points Construction Problem——构造 2020牛客暑期多校训练营(第三场)E-Two Matchings——复杂思维与简单dp 2020牛客暑期多校训练营(第二场)I-Interval——最大流转对偶图求最短路 Educational Codeforces Round 80 D. Minimax Problem——二分+二进制处理 Codeforces Round 606 E. Two Fairs——图论 Codeforces Round 612 (Div. 2) C. Garland——DP
2020牛客暑期多校训练营(第一场)H-Minimum-cost Flow——网络流
Shiroha · 2020-07-14 · via Shiroha白羽的博客

给出一个费用流图,每条边的流量上限相同且不固定。有$q$个询问,每个询问中给出每条边的流量上限(分数,且保证$\leq 1$)。当图中的流量为 $1$ 个单位的时候,求出此时的费用。

首先由于边的流量均为分数($\frac{u}{v}$),而总流量为 $1$ 个单位。我们先扩大$\frac{v}{u}$倍,将每条边的流量固定为 $1$ 个单位,此时流量为 $\frac{v}{u}$ 个单位

考虑在最大流使用 SPFA 查找路径时,当查找到一条路径时,此路径的流量一定为 $1$ (根据上述的设定)。由于采用的本身便是最短路查找路径,得到的路径的费用为当前网络图下的最低费用。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
#define ll long long

const int maxn = 100; //点数
const int INF = 0x3f3f3f3f;

struct Edge {
int from, to, cap, flow, cost;

Edge(int u, int v, int c, int f, int cc)
: from(u), to(v), cap(c), flow(f), cost(cc) {}
};

vector<ll> res;

struct MCMF {
int n, m;
vector<Edge> edges;
vector<int> G[maxn];
int inq[maxn]; //是否在队列中
int d[maxn]; //bellmanford
int p[maxn]; //上一条弧
int a[maxn]; //可改进量
void init(int n) {
this->n = n;
for (int i = 0; i <= n; ++i) G[i].clear();
edges.clear();
}

void addEdge(int from, int to, int cap, int cost) {
edges.emplace_back(Edge(from, to, cap, 0, cost));
edges.emplace_back(Edge(to, from, 0, 0, -cost));
m = int(edges.size());
G[from].emplace_back(m - 2);
G[to].emplace_back(m - 1);
}

bool spfa(int s, int t, int &flow, ll &cost) {
for (int i = 1; i <= n; ++i) d[i] = INF;
memset(inq, 0, sizeof(inq));
d[s] = 0;
inq[s] = 1;
p[s] = 0;
queue<int> q;
a[s] = INF;
q.push(s);
while (!q.empty()) {
int u = q.front();
q.pop();
inq[u] = 0;
for (int i = 0; i < int(G[u].size()); ++i) {
Edge &e = edges[G[u][i]];
if (e.cap > e.flow && d[e.to] > d[u] + e.cost) {
d[e.to] = d[u] + e.cost;
p[e.to] = G[u][i];
a[e.to] = min(a[u], e.cap - e.flow);
if (!inq[e.to]) {
q.push(e.to);
inq[e.to] = 1;
}
}
}
}
if (d[t] == INF) return false;
flow += a[t];
cost += (ll) d[t] * (ll) a[t];
res.push_back(d[t]);
for (int u = t; u != s; u = edges[p[u]].from) {
edges[p[u]].flow += a[t];
edges[p[u] ^ 1].flow -= a[t];
}
return true;
}

int MincostMaxflow(int s, int t, ll &cost) {
int flow = 0;
cost = 0;
while (spfa(s, t, flow, cost));
return flow;
}
} mcmf;

虽然 $1 \leq u, v \leq 1e9$,可能需要循环遍历 $1e9$ 次才能出结果,但是由于图中每条边的容量都是 $1$,所以对于每一条路径,它只有被完全占用和完全空闲两种状态,而边数只有 $100$ 条,即整个网络的最大流量只能是 $100$ 个单位,即最多遍历 $100$ 次,则复杂度不超过 $100 1e5 = 1e7$,当 $MAX_FLOW u < v$ 时,则无解,输出 $-1$

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
#include <bits/stdc++.h>

using namespace std;

#define ll long long

const int maxn = 100; //点数
const int INF = 0x3f3f3f3f;

struct Edge {
int from, to, cap, flow, cost;

Edge(int u, int v, int c, int f, int cc)
: from(u), to(v), cap(c), flow(f), cost(cc) {}
};

vector<long long> res;

struct MCMF {
int n, m;
vector<Edge> edges;
vector<int> G[maxn];
int inq[maxn]; //是否在队列中
int d[maxn]; //bellmanford
int p[maxn]; //上一条弧
int a[maxn]; //可改进量
void init(int n) {
this->n = n;
for (int i = 0; i <= n; ++i) G[i].clear();
edges.clear();
}

void addEdge(int from, int to, int cap, int cost) {
edges.emplace_back(Edge(from, to, cap, 0, cost));
edges.emplace_back(Edge(to, from, 0, 0, -cost));
m = int(edges.size());
G[from].emplace_back(m - 2);
G[to].emplace_back(m - 1);
}

bool spfa(int s, int t, int &flow, ll &cost) {
for (int i = 1; i <= n; ++i) d[i] = INF;
memset(inq, 0, sizeof(inq));
d[s] = 0;
inq[s] = 1;
p[s] = 0;
queue<int> q;
a[s] = INF;
q.push(s);
while (!q.empty()) {
int u = q.front();
q.pop();
inq[u] = 0;
for (int i = 0; i < int(G[u].size()); ++i) {
Edge &e = edges[G[u][i]];
if (e.cap > e.flow && d[e.to] > d[u] + e.cost) {
d[e.to] = d[u] + e.cost;
p[e.to] = G[u][i];
a[e.to] = min(a[u], e.cap - e.flow);
if (!inq[e.to]) {
q.push(e.to);
inq[e.to] = 1;
}
}
}
}
if (d[t] == INF) return false;
flow += a[t];
cost += (ll) d[t] * (ll) a[t];
res.push_back(d[t]);
for (int u = t; u != s; u = edges[p[u]].from) {
edges[p[u]].flow += a[t];
edges[p[u] ^ 1].flow -= a[t];
}
return true;
}

int MincostMaxflow(int s, int t, ll &cost) {
int flow = 0;
cost = 0;
while (spfa(s, t, flow, cost));
return flow;
}
} mcmf;

int n, m;

void solve() {
ios::sync_with_stdio(false);
cin.tie(0);
cout.precision(10);
cout << fixed;
while (cin >> n >> m) {
mcmf.init(n + 10);
res.clear();
for (int i = 0; i < m; ++i) {
int u, v, f;
cin >> u >> v >> f;
mcmf.addEdge(u, v, 1, f);
}
ll cost = 0;
ll mf = mcmf.MincostMaxflow(1, n, cost);
int q;
cin >> q;
while (q--) {
ll u, v;
cin >> u >> v;
if (mf * u < v) {
cout << "NaN" << '\n';
continue;
}
ll sum = 0;
ll ans = 0;
for (auto item : res) {
if (sum + u <= v) {
sum += u;
ans += item * u;
} else {
ans += (v - sum) * item;
break;
}
}
ll g = __gcd(ans, v);
cout << ans / g << '/' << v / g << '\n';
}
}
}

signed main() {
ios_base::sync_with_stdio(false);
cin.tie(nullptr);
cout.tie(nullptr);
#ifdef ACM_LOCAL
freopen("in.txt", "r", stdin);
freopen("out.txt", "w", stdout);
int test_index_for_debug = 1;
char acm_local_for_debug;
while (cin >> acm_local_for_debug) {
if (acm_local_for_debug == '$') exit(0);
cin.putback(acm_local_for_debug);
if (test_index_for_debug > 20) {
throw runtime_error("Check the stdin!!!");
}
auto start_clock_for_debug = clock();
solve();
auto end_clock_for_debug = clock();
cout << "Test " << test_index_for_debug << " successful" << endl;
cerr << "Test " << test_index_for_debug++ << " Run Time: "
<< double(end_clock_for_debug - start_clock_for_debug) / CLOCKS_PER_SEC << "s" << endl;
cout << "--------------------------------------------------" << endl;
}
#else
solve();
#endif
return 0;
}